Live pricing on OneInfer
Pulled from the OneInfer catalog (OpenRouter and direct provider routes) at request time — rates can change. Capture the timestamp with any benchmark or production decision.
| Model | Provider | Input $/1M tokens | Output $/1M tokens |
|---|---|---|---|
| GLM-5.3 | zai | $140.00 | $440.00 |
| GLM-5.3 | zai | $15.00 | $50.00 |
| GLM-5.3 | zai | $37.00 | $125.00 |
| GLM-5.3 | akashml | $117.00 | $396.00 |
| GLM-5.3 | novita | $15.00 | $50.00 |
| GLM-5.3 | together_ai | $15.00 | $50.00 |
| Claude Fable 5 | anthropic | $1000.00 | $5000.00 |
| Claude Fable 5 | anthropic | $1000.00 | $5000.00 |
Decision snapshot
| Decision factor | GLM-5.3 | Claude Fable 5 |
|---|---|---|
| Independent intelligence score | 60 | 62 |
| First-party input price / 1M tokens | $1.40 | $10.00 |
| First-party output price / 1M tokens | $4.40 | $50.00 |
| Context window | 1M tokens | 1M tokens |
| Maximum output | 128K tokens | 128K tokens |
| Input modalities | Text | Text and images |
| Model access | API; weights announced for release | Managed APIs; proprietary weights |
How we keep this comparison honest
- Match the tested modes: the independent table compares GLM-5.3 at max effort with Claude Fable 5 at adaptive max effort on Artificial Analysis Intelligence Index v4.1.1.
- Count Fable's fallback: the published Fable result allows Opus 4.8 fallback, so we label it explicitly instead of presenting the score as a Fable-only base-model result.
- Keep independent and vendor evidence separate: Artificial Analysis results support the main verdict; Z.ai coding scores appear in their own vendor-reported table.
- Compare total task economics, not token price alone: GLM-5.3 has much lower list prices, but used 170M evaluation output tokens versus Fable 5's 83M.
- Freeze time-sensitive facts: prices, capabilities, and availability reflect checks made on August 27, 2026; announced weight releases and provider terms must be rechecked before deployment.
Independent performance and efficiency
Artificial Analysis tested reasoning variants at maximum effort. Its Fable 5 result includes Anthropic's Opus 4.8 fallback behavior, so the score represents the deployed Fable experience rather than an isolated base-model run.
| Artificial Analysis metric | GLM-5.3 (max) | Claude Fable 5 (max, with fallback) |
|---|---|---|
| Intelligence Index v4.1.1 | 60 | 62 |
| Cost per Intelligence Index task | $0.68 | $3.14 |
| Output speed | 84.1 tokens/s | 70.9 tokens/s |
| Total evaluation output tokens | 170M | 83M |
How to read the independent results
- Fable 5 leads the composite intelligence score by two points.
- GLM-5.3 cost about 78% less per evaluated task and generated output faster in this test.
- GLM-5.3 used roughly twice as many output tokens across the evaluation, so lower token prices do not guarantee lower token counts.
- Re-test both models on your prompts because a two-point composite gap does not predict every workload.
Ready to test the workflow?
Create account & add creditsVendor-reported coding benchmarks
Z.ai reports the following results using the same named benchmark rows. These figures are useful directional evidence, but they are vendor-reported and should not be presented as independent verification.
| Benchmark | GLM-5.3 | Claude Fable 5 |
|---|---|---|
| Terminal-Bench 2.1 | 88.2 | 88.0 |
| Terminal-Bench 3.0 | 28.3 | 33.7 |
| DeepSWE v1.1 | 66.9 | 69.7 |
Capabilities and API behavior
| Capability | GLM-5.3 | Claude Fable 5 |
|---|---|---|
| Reasoning | Always enabled; low, high, and max effort | Adaptive reasoning always enabled |
| Tool use | Function calling and structured output | Tool use, code execution, memory, and programmatic tool calling |
| Vision | Not supported by the base GLM-5.3 model | Supported |
| Safety fallback | No equivalent behavior documented | Some requests may refuse or fall back to another Claude model |
Deployment and data-control tradeoffs
Z.ai announced that GLM-5.3 weights would be released on August 28, 2026, after safety evaluation and hardening; at this page's August 27 verification point, they were not yet published. Claude Fable 5 is proprietary and available through Anthropic and supported cloud platforms. Anthropic documents mandatory 30-day retention for Fable 5 and no zero-data-retention option, which may affect regulated workloads.
Which model should you choose?
| Workload or priority | Recommended starting point | Why |
|---|---|---|
| Highest measured general capability | Claude Fable 5 | It leads the independent index and the harder vendor-reported coding rows. |
| Budget-sensitive coding agents | GLM-5.3 | Its first-party token prices and measured task cost are much lower. |
| Image or screenshot understanding | Claude Fable 5 | Fable supports image input; base GLM-5.3 is text-only. |
| Long-context text analysis | Test both | Both expose a 1M-token context window and 128K maximum output. |
| Future self-hosting or model control | GLM-5.3 | Z.ai has announced an open-weight release; verify that the weights and license are available before deployment. |
| Strict zero-retention requirement | Review alternatives | Fable 5 requires 30-day retention; confirm GLM provider terms before sending data. |
Frequently asked questions
Can I try GLM-5.3 before integrating it?
Use the OneInfer GLM-5.3 launcher to open a prepared prompt in the authenticated playground. Availability is checked against the current model catalog.
How should I treat benchmark claims?
Check the provenance label and harness version. Vendor-reported and independently verified results are deliberately shown as different evidence classes.
Is GLM-5.3 cheaper than Claude Fable 5?
At first-party list prices verified on August 27, 2026, yes. GLM-5.3 is $1.40 per million input tokens and $4.40 per million output tokens; Fable 5 is $10 and $50 respectively. Provider prices and discounts may differ.
Do both models support one million tokens of context?
Yes. Both vendors document a 1M-token context window and a maximum output of 128K tokens. Effective accuracy over very long inputs still needs workload-specific testing.
Can GLM-5.3 process images?
The base GLM-5.3 model accepts text only. Claude Fable 5 supports text and image input. Do not confuse GLM-5.3 with separate multimodal GLM variants.
Put GLM-5.3 to work
Fund a controlled evaluation, start with a prepared prompt, and measure quality and cost on your own workload.